The nucleosome is a histone DMAcomplex that folds 147 basepair of DMA into a 1.7 turns of a left-handed superhelix. The core histones can fold any sequence of DNA into a nucleosome. However, the affinity of the histone core for different sequences varies so that it will assume a preferred location on a length of DNA that is greater than 147bp. This phenomenon is called nucleosome positioning. Since nucleosomal DNA is significantly deformed compared to its conformation free in solution, the flexibility and any intrinsic bends in DNA affect nucleosome positioning. Idealized models of the nucleosome have been successfully employed to rationalize relationships between DNA sequence and nucleosome positioning. In these models the nucleosome superhelix is assumed to have an ideal conformation. For a twisted fiber, like DNA, that possess two different bend stiffnesses there will be a sinusoidal variation in bending such that bends in the softer direction vary with the twist in the fiber. While such models are useful in developing general rules for identifying positioning signals they have failed for a number of sequences. Moreover such models have not been able to accurately predict experimental measures of relative binding affinities. There are now a number of high-resolution x-ray crystallographic structures available enabling us to replace the assumed ideal superhelix with actual conformational data. Unfortunately the x-ray structures all have essentially the same sequence of DNA. Our long-term objective is to overcome this shortcoming using all atom molecular modeling and molecular dynamics simulations and to develop a more accurate model of nucleosome stability. For this purpose we will simulate nucleosomes and free DNA with sequences that span the range of known binding affinities. Conformational and dynamical properties will be extracted from our simulations and incorporated into a statistical mechanics based model of nucleosome positioning. Our methods will then be applied to investigate the effects of DNA chemical modifications and hormone receptor binding on nucleosome stability. Two positioned nucleosomes will be studied. One is from the Mouse Mammary Tumor Virus (MMTV) promoter. The other is from the promoter for the a isoform of the estrogen receptor (ERa). These receptors are known to bend DNA. Both function according to the hormone response mechanism, which regulates a number of physiologic processes. Both are linked to cancer treatment and diagnosis.